INDEX

Explanations

phrases related to bias and its implications in algorithms and decision-making

oai_token-act-pair · gpt-4o-mini Triggered by @bot

therefore

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 betweenstory

-0.84

intios

-0.63

ArrowToggle

-0.63

 Biôgrafia

-0.60

 autorytatywna

-0.57

 مرئيه

-0.57

 محفوظة

-0.57

 deſſen

-0.56

 Neighbors

-0.56

сылкі

-0.55

POSITIVE LOGITS

 therefore

0.34

代

0.28

Therefore

0.28

 making

0.28

 sitä

0.27

 plegable

0.27

 Therefore

0.27

 donc

0.27

 this

0.27

 MAKING

0.27

Activations Density 0.171%

phrases related to bias and its implications in algorithms and decision-making

therefore

No Comments

No Known Activations